Pixel-based representations of data allow for large data sets to be represented in a graphically efficient manner. For example, time series data may be represented in a two dimensional (2D) pixel calendar tree that subdivides the data into calendar elements. In this example, each pixel within the representation may represent a portion of the time series data for a period of time (e.g., hour, day, week, etc.). Such pixel-based representations may be used to determine patterns and anomalies in the data sets.